模糊推理方法在股市预测中的应用-《华侨大学学报（自然科学版）》

文章信息/Info

Author(s):: Hu Yilang; Dept. of Manag. Info. Sci., Huaqiao Univ., 362011, Quanzhou

Keywords:: fuzzy inference; stock price prediction; turnover; relative strength index

摘要:: 以日成交金额和 15日相对强弱指标的绝对量和增量共 4个变量作为影响股指走势因素,应用模糊推理方法对沪市中短期上升行情进行预测,结果较为满意 .模糊推理方法对每一种影响因子 (xi)与输出因子 (y),统计其在每一分级中占的次数,取两者次数的比值即为分级模糊关系值,从而得到模糊关系矩阵 R(xi,y) .再取 R(xi,y)每行最大值相加,可得到 4个数值 .以每个数值除以 4个数值之和,即为每个影响因子权重 wi.最后,建立预测模型 y=w R* .R*是由所预测行情的 4个变量决定的矩阵

Abstract:: Four variables,absolute amount and increment of daily turnover and relative strength of 15 days,are taken as factors influencing the trend of share exchange.The method of fuzzy inference is applied to predicting medium term and short term bull market of shanghai share exchange,with fairly satisfactory result.The method is adopted in proper steps.Firstly,various influencing factors are graded,and their weights and fuzzy relationship are solved statistically.Secondly,the orders of every influencing factors in every grade ( x i) and the output factor in it ( y ) are counted up,the ratio of their orders is just the graded fuzzy relationship from which the fuzzy relational matrix R( x 1,y ) can be obtained.Thirdly,to add up the maximum values in every line of R( x 1,y ),four numerical values can be obtained; and to divide every numerical value by the summation of four ones,the weight of every influencing factor can be obtained.Finally,to set up a prediction model y=w R *,where R * is the matrix decided by four variables for predicting market.